The first phase of GVHD started already before the actual HSCT with the conditioning regime. This treatment depletes the patient’s own bone marrow cells to allow engraftment of the donor’s stem cells and prevent their rejection. It also aims at clearing any remaining pathological cells in the patient. The myeloablation is either done by total-body irradiation or high-dose chemotherapy. However, neither strategy is targeted and causes also injury to other tissues, particularly the gut mucosa, skin, and liver. The damaged cells in response secrete a massive amount of cytokines, especially interleukin (IL)-1, IL-6, interferon (IFN)-γ, and tumor necrosis factor (TNF)-α [11].

Several polymorphisms for genes of these cytokines have been associated with an increased risk for GVHD. While polymorphisms for TNF-α, IFN-γ, IL-6, IL-10, and a natural occurring IL-1 receptor antagonist were associated with severity of acute GVHD [12–15], IL-6 also correlates with the incidence of chronic GVHD [16–18].

Not only cytokines but also other danger signals contribute to the proinflammatory environment. One molecule that is released from dying cells and has been implicated in the pathogenesis of GVHD is ATP [19]. It can itself induce the secretion of the proinflammatory cytokines IFN-γ, TNF-α, and IL-6 and activate antigen-presenting cells (APC). Also the ATP receptor P2X7 is upregulated during GVHD, leaving cells more susceptible to the effects of ATP and creating a positive feedback loop. Thus, blocking of P2X7 as well as reduction of ATP concentration can improve GVHD symptoms. Even increased numbers of regulatory T (Treg) cells, which are known to control GVHD, are detected. Notably, Treg cells are also one of the main populations expressing the ATP-hydrolyzing enzyme CD39 on their surface [20] as they are particularly sensitive to ATP [21].

In the intestine, the tissue damage leads to the release of microbial lipopolysaccharides (LPS) which is recognized by Toll-like receptors (TLR) on APC. Polymorphisms in genes for several of these TLR, namely, TLR4, TLR7, and TLR9, were described to influence the incidence and severity of GVHD [22–24]. Furthermore, flagellin, an agonist for TLR5, was able to reduce GVHD in a mouse model [25]. Although the TLR/APC axis undoubtedly plays an important role in the onset of GVHD, it also develops without intact TLR signaling [26].

The presence this plethora of cytokines and danger signals present at these sites of acute inflammation leads to the activation of APC. These comprise dendritic cells (DC), monocytes, macrophages, and B cells which present antigens to T lymphocytes and activate these. APC and other local cells upregulate the MHC (major histocompatibility complex) and co-stimulatory and adhesion molecules [27] and thereby set the stage for the next phase of GVHD.

The second phase is initiated when donor stem cells are transferred into this proinflammatory setting. They are attracted by the presence of cytokines, chemokines, and adhesion molecules at the sites of tissue injury. Here they come into contact with the activated host APC that present both MHC molecules and self-antigens. Any self-antigen derived from a protein that is different between donor and recipient can give rise to minor histocompatibility antigens (MiHA). Allogeneic donor cells recognize disparities in MHC for HLA mismatched donor/patient pairs or in MiHA with their T-cell receptor (TCR) [28]. In addition, also non-hematopoietic host tissue cells [29, 30] and donor APC [31, 32] can be involved in the presentation of host antigens. As all of these cell types are involved in the presentation of alloantigens, targeting only one cell type is not sufficient in abrogating GVHD [33, 34].

In addition to TCR engagement, alloreactive T cells require co-stimulatory signals for their activation; these are also provided by activated APC. The main pathway for this co-stimulation involves the engagement of CD28 on T lymphocytes by its B7 ligands CD80 and CD86 on APC [35]. Also CD40 and its ligand CD40L play a role in providing positive co-stimulation for T cells [36].

After activation by antigen recognition plus co-stimulation, allogeneic T cells differentiate into effector cells [37] and secrete TH1 cytokines, mainly IFN-γ, IL-2, and TNF-α, as well as the TH17 cytokine IL-17 [38], and thereby execute phase III of GVHD.

The clinical symptoms of acute GVHD are mainly caused by the action of T cells. It was shown that only the naïve pool of donor T cells is responsible for the development of GVHD but not transplanted effector and central memory T cells [39]. Other important cell subsets responsible for the tissue damage in the effector phase of GVHD are natural killer (NK) cells, macrophages, and neutrophils [40]. Also B cells were reported to play a role [41].

Alloreactive CD8⁺ T cells use two main mechanisms for their cytotoxic action, both of which induce the activation of the caspase cascade in target cells by direct cell contact. One pathway induces the release of perforin and granzymes which trigger cell lysis [42]. The other mechanism requires engagement of the death receptor, Fas, with the Fas ligand (FasL) leading to apoptosis [43].

Tissue damage is also induced indirectly by the cytokines present in abundance from the preconditioning (TNF-α, IFN-γ and IL-1).

These cytokines themselves trigger the release of a surge of additional cytokines from allogeneic donor cells in a self-amplifying cycle that can lead to a cytokine storm [44] with deleterious effects on the patient.

TNF-α is the key cytokine involved in all three phases of the pathogenesis of GVHD [45]. It was reported as an independent predictor of GVHD as it correlates with its severity, time of onset, and mortality when measured with the help of the surrogate marker TNF receptor 1 (TNFR1). TNFR1 binds TNF-α and is shed into the plasma where it can easily be measured [46–49]. Another member of the tumor necrosis factor receptor family, CD30, has been described as a marker for GVHD. Higher levels of CD30 were found both in plasma and on CD8⁺ T cells of patients with acute GVHD [50]. A more comprehensive combination of the four biomarkers IL-8, hepatocyte growth factor, TNFR1, and IL-2 receptor-α was suggested to predict the occurrence of GVHD [51].

Natural suppression of GVHD is conferred by Treg cells. These cells regulate both innate and adaptive immune responses and confer peripheral tolerance by suppressing activated T cells [52, 53]. Their modes of action include a variety of mechanisms involving both soluble factors and direct cell-cell contact [54, 55]. In mouse models, depletion of Treg cells exacerbated GVHD, while infusion of these cells improved its outcome [56, 57]. However, in the overwhelming inflammation during GVHD, the balance of the immune system is grossly shifted toward the proinflammatory response and apparently leaves Treg cells insufficient. Nevertheless, Treg cells have been associated with the outcome of GVHD [57, 58]. Even specific markers in the transcriptome of Treg cells were able to predict both acute and chronic GVHD, among these were activation markers (PIP5Kγ, FAS, CD44, CD69), cyclins, and the transcription factor Eos which is important for the suppressive capacity of Treg cells [59].

The undesirable graft-versus-host syndrome is closely intertwined with the graft-versus-tumor (GVT) response, which is part of the curative effect of stem cell transplantation especially in patients with malignancies. T cells from the graft not only recognize alloantigens but also tumor antigens and are thus able to act against leukemia or tumor cells that escaped the conditioning regime. Also NK cells [60–62] and NKT cells were implicated in the GVT activity [63]. The greater the MHC mismatch between donor and patient, the stronger the GVT response is. This explains the advantage of allogeneic over autologous HSCT and the higher relapse rate of patients receiving HSCT from a related donor [1]. Especially after non-myeloablative pretreatment of the patient, GVT is indispensable to clear any remaining malignant cell. The strong general immunosuppression as well as T-cell depletion to prevent GVHD also inhibits GVT [64]. Thus, the challenge is to find a balance between the negative impact of GVHD and the positive effects of GVT. The incidence of GVHD is considered as an indication of an ongoing GVT activity and was shown to correlate with a reduced risk of relapse [65, 66]. An option to treat cancer relapse after HSCT is donor lymphocyte infusion which has a high risk of GVHD but also a highly effective GVT response [67].

Approaches to separate GVHD from GVT are difficult as both share similar mechanisms. However, attempts are being made to interfere with differences in the organ-specific proinflammatory environment, specific cells involved, and selective migration of effector cells to the target tissues [68].

Without any preventive treatment, acute GVHD would occur in practically every patient receiving allogeneic HSCT. Even with preemptive general immunosuppression, acute GVHD is still a major complication after HSCT. The standard treatment includes corticosteroids, cyclosporine A, and methotrexate (MTX). The combination of these has proven to be most effective [2]. Most novel approaches for the treatment of GVHD discussed below are also tested for its prevention.

A non-pharmacological approach to prevent the occurrence of GVHD is depletion of T cells, either ex vivo from the graft or in vivo. This depletion can be achieved by different methods, like anti-thymocyte globulin (ATG) or monoclonal antibodies (mAbs) directed against CD4, CD5, CD8, and/or CD52 (Campath/alemtuzumab) [69, 70]. Although depletion of lymphocytes was shown to reduce GVHD, it also increases the risk of graft rejection by the recipient’s T cells and of recurrence of the underlying disease as also GVT responses are diminished. It needs to be shown whether the attempts to selectively deplete alloreactive T cells can provide a tool to reduce the incidence and severity of GVHD.

Even with these comprehensive measures of prevention, between 10 and 50 % of patients still develop forms of GVHD severe enough to require additional treatment. The first-line therapy remains administration of steroids, in particular methylprednisolone. However, the high doses necessary to abrogate the ongoing immune reaction lead to a generalized immunosuppression with the risk of opportunistic infections [2].

Other routine medications like the calcineurin inhibitor cyclosporine A inhibit T-cell activation and block IL-2 transcription [71]. The macrolide antibiotic tacrolimus or FK-506 also interferes with IL-2 production [72, 73] but appears to have less toxic side effects [74]. A different pathway is targeted by mycophenolate mofetil, which inhibits activation and proliferation of lymphocytes by blocking their purine synthesis [75]. These standard drugs, even in combination, are not sufficient for all patients and have serious side effects in addition to the immunosuppression [76–78]. Often renal and other toxicities limit their use.

Another approach is the depletion of T lymphocytes by the polyclonal anti-thymocyte globulin (ATG). A meta-analysis of six prospective randomized controlled trials found that ATG was able to lower incidence of grade II–IV GVHD with a stronger effect seen for rabbit compared to equine ATG but no impact on overall GVHD, survival, or relapse [79]. Higher doses of ATG can further decrease incidence of GVHD, but with a higher risk of serious infections [80]. A more recent study with a different ATG preparation could confirm the reduction in GVHD and reported a reduction in tumor relapse and infections [81]. Also a reduction of chronic GVHD was reported [82].

An interesting therapeutic modality is the extracorporeal photochemotherapy (ECP), a combination of a photosensitizing drug with exposure to UVA light. This has been demonstrated to inactivate lymphocytes although the mechanism still needs to be elucidated. However, in several clinical trials, symptoms of both acute and chronic GVHD were improved without negatively affecting GVT or relapse incidence [83, 84].

Also statins were shown to have an immunomodulatory effect by reducing cytokine secretion and T-cell activation [85]. In a mouse model, they reduced GVHD incidence [86]. A correlation study in patients could attribute a decreased risk for acute GVHD to the use of statins by the donor [87]. The risk for chronic GVHD was diminished if the recipient took statins [88].

A novel mechanism is targeted by the proteasome inhibitor bortezomib which is approved for treatment of myeloma. In a mouse model, early administration promoted the apoptosis of alloreactive T cells and thus prevented acute GVHD [89]. These results could be corroborated in a small clinical trial which also suggested an improvement of immune reconstitution [90]. However, timing is crucial as delayed and continued administration enhanced GVT but also aggravated GVHD symptoms [89].

GVHD is well known as a multifactorial process with a variety of players involved, and most of these have been targeted in mouse models. However, not all could be transferred into clinical applications. Therefore, this chapter will focus on approaches for which clinical trials have been reported.

Hematopoietic stem cell transplantation is often the last resort for patients with malignancies although the long-term outcome is still disappointing. One major complication is GVHD, a fulminant inflammatory response that is difficult to treat with conventional immunosuppression. New approaches target the different pathways involved in GVHD. Results of first clinical trials are promising, but no therapy is efficient for all patients. Heterogeneous patient cohorts and concomitant treatments make it difficult to compare results from different studies. The networks involved in the pathogenesis of GVHD are complimentary, and it is unlikely that targeting only one pathway will be sufficient. It is also possible that certain pathways are more important in one patient group than in others. In addition, genetic differences contribute to this heterogeneity and might be responsible for discrepancies observed in many trials. A more thorough selection of patients for specific therapies might give clearer results in the future. It might also prove helpful to identify relevant pathways to provide a more tailored therapy for individual patients which would not only improve efficacy of the therapy but also help to reduce side effects which are a huge problem with current medications. An additional challenge is to specifically prevent GVHD without affecting general immunity and the GVT effect. Eventually, combination of different therapeutic strategies and more individualized treatment of patients could improve the outcome and therefore boost the success story of HSCT.